A Short-Term View of Protein Sequence Evolution from Salmonella.

Abstract

Much of the study of protein sequence evolution is based on sequence changes inferred to have occurred in nature. The sequences compared for this purpose are usually sufficiently distant that purifying selection has had nearly its full effect and most of the changes inferred have been exposed to a variety of conditions. Here, I make use of large numbers of Salmonella genome sequences to study changes known to be of very recent origin because they are inferred from comparison of very closely related sequences. The effects of purifying selection are weak yet discernible on this short timescale: the ratio of nonsynonymous to synonymous changes is smaller than expected under selective neutrality, but only slightly so. Essential genes have lower rates of nonsynonymous change, as they do on a longer timescale, but much more of this association remains after controlling for expression level. Positive selection for nonsynonymous change is inferred for 151 genes. For nearly half of these, this is attributable to selection for loss of function. Other forms of positive selection inferred include selection for amino acid changes that make enzymes less sensitive to antibiotics and selection for activating changes to proteins involved in transcriptional regulation. Positively selected variants of many genes are likely favored only under unusual conditions and disfavored in the long term, making detection of the positive selection with more distant comparisons difficult or impossible. The short-term view provided by close comparisons complements the long-term view obtained from more distant comparisons such as those between species.